Volume optimization, normal surfaces, and Thurston's equation on triangulated 3-manifolds (Q360072)

To a triangulated 3--manifold are associated various systems of equations. One of the most studied is Thurston's gluing equation, which is given by a system indexed by the edges of the triangulation, to each \(e\) being associated the single equation: \[ \prod_i z_i = 1 \tag{Thurston} \] where the product is over all tetrahedra \(T_i\) containing \(e\), and \(z_i\) is the complex parameter associated to \((T_i,e)\) by realizing \(T_i\) as an ideal tetrahedron in hyperbolic three-space. Another noticeable equation is Haken's normal surface equation, the solutions to which yield certain embedded surfaces in the manifold. The aim of this paper is to explore the relation of the existence of solutions for Thurston's and Haken's equations to the existence of solutions with special properties of another equation (for which solutions are always known to exist). More precisely, the author defines (generalizing previous constructions by A. Casson, I. Rivin and M. Lackenby) the notion of an \(\mathbb{S}^1\)-angle structure on a triangulation of a 3-manifold \(M\) (Definition 2.3). The space of these structures on a given \(M\) has the structure of a smooth manifold, and one can define the volume of such a structure, which determines a volume on each simplex by a natural analogue of the volume of a hyperbolic ideal simplex, and this yields a continuous function on this space. The main result of the paper (Theorem 1.1) says that a smooth maximum of this function produces a solution to the generalized Thurston equation (where one asks for the left-hand side of the above equation (ref {Thurston}) only to be \(\pm 1\)). The alternative (if the given maximum of the volume function is nonsmooth) is that it supports a solution to the normal surface equation with certain additional properties, and by a previous result of S. Tillmann and the author this results in severe topological constraints on \(M\) (listed in Theorem 1.2 in the paper). The author further conjectures that if all maxima are smooth then the triangulation actually supports a genuine solution to the above equation (ref {Thurston}). To give an idea of the strength of his conjecture he notes that, by a previous result of H. Segerman and S. Tillmann, this implies (if the triangulation has a single vertex) that \(M\) cannot be simply connected, and thus that the Poincaré conjecture would follow from his.